Table 2

Case study of the dry stone wall principle applied to an intervention study

Intervention consideredMain research questionKey principles of approach takenExamples of key methods and findings
‘A new (…) section of (…) motorway (…) running through predominantly deprived neighbourhoods (…) with associated changes to the urban landscape’44‘What are the individual, household and population impacts of a major change in the urban built environment on travel and activity patterns, road traffic accidents and well-being?’Looking beyond interventions
‘Numerous health-related claims were made for and against the new motorway (…) We summarised these (…) as two equally valid, competing, testable, overarching hypotheses (…) expressed in the form of vignettes of two alternative extreme cases, a ‘virtuous spiral’ and a ‘vicious spiral’ (…) using the developing situation (…) to understand more about the positive and negative effects of the changes to the urban landscape’‘Mapping our findings against the key propositions of each vignette, we find—perhaps unsurprisingly—a mixture of confirmatory and disconfirmatory evidence on both sides’
Searching for patterns
‘We sought to build an evidential case for causal inference using multiple sources of data and types of analysis (…) by taking a ‘pragmatic pluralist’ approach to the ‘ragged evidence’ of the natural experimental study (…)’‘The study used a combination of quantitative (cohort, cross-sectional, repeat cross-sectional and interrupted time series) and qualitative (documentary analysis and interview) research methods to evaluate both individual-level and population-level changes in health and health-related behaviour, and to develop a more in-depth understanding of how these changes were experienced and brought about’
Embracing the mess
‘We sought to match patterns of outcomes with patterns predicted by the intervention theory imperfectly captured in these vignettes, searching not for support for a singular overarching hypothesis, but rather for the least implausible explanation of the conditions that may be required to produce or prevent the outcomes of interest’‘Our evidence points to two critical functions—connecting and separating—that constitute two sides of the same coin and are both evoked by the same intervention in different ways for different people (…) The overarching hypothesis with which our data are most consistent is that new transport infrastructure is more likely to benefit more people when it connects people with their social and physical surroundings—broadly defined—more than it separates them, and when people are protected from its harmful environmental impact by distance or other effective mitigation measures’
  • Emphases added.